Carbonylation of Hydridopentacyanocobaltate (111)

plexes [CO(S~CSR)~(SR)I, analogous to those of iron-. (111). These can be prepared by the thermal elimina- tion of carbon disulfide from CO(S~CSR)~. T...
1 downloads 11 Views 355KB Size
1575

does not make a geometric adjustment to accommodate the extra electron. Slight changes in geometry brought about by differences in amino acid composition could alter the relative energies of the reduced form of the proteins and thus also provide an explanation for the differences in redox potential of, for example, spinach ferredoxin (Eo' = -0.430) and putidaredoxin (EO' = -0.235 V).4a Certain features of the proposed explanation which are subject to experimental testing are currently being investigated.

metal-metal bond. The differences between this structure (11) and that of the iron complex of identical composition (I) are a consequence of their different electronic structures (Figure 1). (3) In solution, the cobalt dimers have structure 11, as determined by pmr studies. Paramagnetic anisotropy effects are shown to be of the correct order of magnitude to account for the large upfield shifts of the methylenic protons attached to bridging sulfur atoms. (4) The monomeric complexes C O ( S ~ C S R )exhibit ~ reversible one-electron reduction waves, extending the number of known compounds containing only four-membered chelate rings that show Summary reversible electrochemical behavior. ( 5 ) The electrochemical behavior of sulfur-bridged dimers may be inThe main findings of this work may be summarized fluenced by the geometric constraints on, and metalas follows. (1) There exists a series of binuclear commetal bonding within, the M2S2 moiety. Such an plexes [CO(S~CSR)~(SR)I, analogous t o those of ironargument has been introduced to rationalize the low (111). These can be prepared by the thermal eliminaredox potentials of the Fe2S2non-heme iron proteins. tion of carbon disulfide from CO(S~CSR)~. The reaction is first order in monomer, with the rate-determining Acknowledgments. We are grateful to the National step being cleavage of the CoS2C-SR bond. (2) The Institutes of Health for support of this research under solid state structure of [ C O ( S ~ C S C ~ H ~ ) ~ ( S conC ~ H ~ ) Grant ]~ No. G M 16449, to Mr. Robert Winograd for tains bridging ethyl mercaptide groups in the anti conexperimental assistance, and to Dr. T. Moss for stimufiguration, four terminal thioxanthate ligands, and no lating discussions. S. J. L. also thanks the Alfred P. Sloan Foundation for a Research Fellowship (1968and W. R. Dunham, Sfruct. Bonding (Berlin), 8, 1 (1970), for one pos1970). sible orbital representation of this state.

Carbonylation of Hydridopentacyanocobaltate (111) Giovanni Guastalla, Jack Halpern,* and Marijan Pribanib Contribution f r o m the Department of Chemistry, The University of Chicago, Chicago, Illinois 60637. Received July 10, 1971 Abstract: Carbon monoxide was found to react with hydridopentacyanocobaltate(II1) in aqueous solution accord2CN- HzO. Kinetic measureing to the stoichiometry: CO(CN)~H~- 2CO OH- -+ CO(CN)~(CO)~~/~~ where k = 6.3 X loe3 M-' sec-' at 23' ments yielded the rate law, - ~ [ C O ( C N ) ~ H ~ -=- ]~[CO(CN)~H~-~[OH-] and 1 Mionic strength. The results are interpreted in terms of the mechanistic sequence, CO(CN)~H~-OHCO(CN)~~- H20 (ks, k-5); Co(CN)S4- + Co(CN)h3- CN- (ks); Co(CN)a3- 2CO + Co(CN)3(C0)z2CN- (fast). In combination with previously determined values of ks and k-5, the present kinetic measurements yield k s = 4.5 X lo3 sec-1.

+

+

+

+

E

xtending our earlier studies on the carbonylation of pentacyanocobaltate(II), we have found that, in the presence of excess base, carbon monoxide reacts with hydridopentacyanocobaltate(II1) in aqueous solution in accord with the stoichiometry described by eq 1. In this paper we describe a study of the kinetics of this reaction and discuss its mechanism. The results of these studies are also of some interest in relation to several other reactions of C O ( C N ) ~ H ~studies -, of which have recently been reported. 2-4 Co(CN)5H3-

+ 2CO + OHCo(CN)3(C0)t2-

+ 2CN- + H2O

+

+

(1)

Experimental Section

+

-

I

(1) J. Halpern and M. Pribanif, J. Amer. Chem. SOC..93. 96 (1971). (2) J. Hanzlik and A. A. VIEek, Inorg. Chem., 8,669 (1969). (3) H. S. Lim and F. C . Anson, ibid., 10, 103 (1971). (4) J. Halpern and M. Pribanib, ibid., in press. .

* +

chloride, sodium cyanide, sodium hydroxide, and sodium chloride (used to adjust the ionic strength) in distilled water with rigorous exclusion of oxygen (achieved by purging with nitrogen which had been passed twice through acidified chromium(I1) chloride solution). Solutions of C O ( C N ) ~ H ~(typically ca. 0.03 M ) were prepared by hydrogenation of corresponding solutions of Co(CN)s3according to eq 2 in a glass-lined Parr Series 4500 stirred autoclave, pressurized with approximately 20 atm of HPas previously de~ c r i b e d . ~Under these conditions the equilibrium of reaction 2 lies sufficiently far to the right (and the reverse reaction following release of the hydrogen pressure is sufficiently slow) that interference from reaction of residual Co(CN)&3- was negligible. The Co(CH)6H3- concentration was determined spectrophotometrically using the 305-nm absorption band (e 610). Carbon monoxide and hydrogen gases were Matheson CP grade. 2Co(CN)5*-

Materials. Solutions of CC(CN)~~- of the desired composition were prepared by dissolving- analytical reagent grade cobalt(I1) ~,

+

+ Hz

2C0(CN)sH3-

(2)

Stoichiometry Measurements. In addition to the spectrophotometric measurements to be described, the stoichiometry of the reaction was confirmed by direct determination of the uptake of CO and the release of CN-. The volume of CO taken up at constant

I

( 5 ) (a) J. Halpern and L. Y . Wong, J. Amer. Chem. SOC.,90, 6665 (1968); (b) J. Halpern and M. Pribanit, Inorg. Chem., 9, 2616 (1970).

Guastalla, Halpern, Pribanid- Carbonylation of Hydridopentacyanocobaltate(III)

1576 pressure as the reaction proceeded was measured using the gasburet apparatus and the procedure described previously.7 The corresponding increase in free C N - was measured by spectral titration with Co2+ a t 970 nm to an end point corresponding to quantitative formation of Co(CN5+. Kinetic Measurements. The kinetics of the reaction were determined by two independent methods, namely (1) measuring the uptake of CO at constant pressure using the gas-buret apparatus and the procedure described previously,6 and (2) monitoring the forma~at 250 nm where tion of C O ( C N ) ~ ( C O ) ~spectrophotometrically CO(CN),(CO)~~has a shoulder (e 8 X IO3) and the other species are virtually transparent. In the CO uptake experiments the initial Co(CN):,H3- concentration was typically ca. 5 X 10-3 M and in the spectrophotometric experiments ca. 1 X lo-' M . The initial concentration (corresponding to saturation with 1 atm of CO) in the M . The spectrophotometric latter experiments was cn. experiments were performed in a Cary 14 spectrophotometer equipped with a thermostatted (k0.2") cell compartment. The reactions were initiated by injecting, with a microsyringe, a calibrated small volume of a standard solution of Co(CN);H3- (typically 0.020 ml of 0.030 M Co(CN);H3-) into a serum-capped spectrophotometer cell filled with a solution containing the desired concentrations of all the other reaction components and thermally equilibrated to the reaction temperature. The reactions were sufficiently rapid so that they could be followed essentially to completion.

Results and Discussion Stoichiometry. The following observations served to confirm the stoichiometry of the reaction described by eq 1. (1) In each case the final spectrum of the product solution was in quantitative agreement with the spectrum of C O ( C N ) ~ ( C O ) ~previously ~characterized as one of the products of the reaction of CO with C O ( C N ) ~ ~ -(2) . ~ ~Gas uptake measurements, made as described earlier, yielded values of 1.8 f 0.1 for the ratio of the number of moles of CO taken up to the number of moles of C O ( C N ) ~ H ~originally present. (3) The ratio of the CN- liberated during the reaction to the initial C O ( C N ) ~ H ~concentration was determined, as described earlier, to be 2.0 f 0.1, in accord with eq 1. (4) While attempts to isolate pure salts of C O ( C N ) ~ ( C O ) ~were ~ - invariably unsuccessful because of accompanying decomposition and disproportionation, other pure cobalt(1) derivatives of the type PR&] (where CO(CN)(CO)~(PR~)~ and K[ CO(CN)~(CO)( PR3 = P(C2Hj)3,P(C6H5)3,etc.), formed by the addition of the corresponding phosphines to the product solutions, could readily be isolated in pure form and fully characterized. A number of such new compounds have now been prepared and their chemistry is being investigated.' Kinetics. Kinetic measurements were performed over the temperature range 13.0-33.2" and over a wide range of solution compositions encompassing the following initial concentration ranges: 9 X 10-5-5 X M C O ( C N ) ~ H ~ 0.20-1.0 -, M OH-, 4.7 X 1.2 X M CO (0.50-1.2 atm), and 9 X 10-j-0.25 M CN-. The ionic strength, adjusted with NaC1, was held constant at 1.0 M . The kinetic results conformed to the rate law described by eq 3. Under the conditions of our measurements, with CO in excess over Co(CN):H3- and with the OH- concentration sufficiently high to remain virtually constant throughout each experiment, the observed kinetic behavior was always pseudo first order

0 03

1

(over at least 3 half-lives) in accord with eq 4, where kobsd = k[OH-1.

k[C o(CN) 6H "I[ OH-]

- d[C O( CN)jH +]/d t =

-d In [Co(CN)jH3-]/dt = kobsd

Journal of the American Chemical Society 1 94:5

(4)

Typical pseudo-first-order rate plots, derived from experiments in which the reactions were followed both spectrally and by CO uptake, are depicted in Figure 1. The results of the kinetic measurements are summarized in Table I where the value of k at constant temperature M-' sec-l at 23'1 is seen to be vir[(6.3 f 0.5) X tually independent of the concentration of C O ( C N ) ~ H ~ - , OH-, and CO and unaffected by added CN-. Measurement of the temperature dependence of k yielded the 1 kcal/mol and activation parameters, AH* = 16 A S * = -17 + 4 eu. Mechanism. The rate law represented by eq 3 is similar in form to the rate laws found for several other reactions of C O ( C N ) ~ H ~including -, the reaction with Fe(CN)63-,4as well as the limiting form of the rate law for the reaction with Hg(CN)z. Both of these reactions have been interpreted in terms of mechanisms involving the rate-determining step depicted by eq 5, followed by subsequent fast reactions of the Co(CN)b 4- ion formed in this step with the substrate, Le., with Fe(CN)c3- or Hg(CN)Z. Co(CN),H3-

+ OH-

ks

Co(CN)h4- f HzO k-

(6) A. J. Chalk and J. Halpern, J . Amer. Chem. SOC.,81, 5846 (1959). (7) J. Bercaw, G. Guastalla, and J. Halpern, Chem. Commim., 1594 (1970).

(3)

(5)

3

However, this simple mechanistic scheme cannot be extended to the present reaction with CO since the value

March 8,1972

1577 Table I. Summary of Kinetic Data" [OH-], M

Temp, "C

0.20 0.29 0.40 0.39 0.50 0.58 0.60 0.68 0.78 0.80 0.87 0.97 1 .oo 0.97 0.97

23.0 23.0 23.0 23.0 23.0 23.0 23.0 23.0 23.0 23.0 23.0 23.0 23.0 13.0 33.2

Methodb SP

Gas SP

Gas SP

SP

SP SP

Gas SP Gas Gas SP

Gas Gas

Co(CN)aC03103k,~,~, sec-'

103k, M-1 sec-1

1.4 2.0 2.4 3.0 3.3 3.9 3.6 4.2 4.8 4.8 5.1

-

a Initial Co(CN)&13-concentrations ea. 5 X 10-3 Min CO uptake experiments and ca. 9 X lo+ M in spectral experiments; [CO] M (1 atm partial pressure) unless otherwise stated; ionic sp = spectral; gas = CO strength adjusted to 1.0 M with NaC1. uptake. Unaffected by variation of [CN-] from 9 x 10-5 to 0.25 M . Unaffected by variation of [CO] from 4.7 X to 1.2 X M (0.5-1.2 atm).

M-I sec-' at 23", of k for this reaction (i.e., 6.3 X 1.0) is significantly lower than the corresponding value of ks (1.6 X lo-' M-I sec-I at 25", p = l.0)4for the reaction with Fe(CN)63- (or with Hg(CN)Z). The activation parameters are also significantly different. Hence, despite the similarity of the forms of their rate laws and the independence of the rate constants in each case of the substrate (Le., CO or Fe(CN)63-) concentration, both reactions cannot proceed through mechanisms characterized by a common uniquely rate-determining step. To explain this behavior we propose a mechanism for the reaction of Co(CN)sH3- with CO (i.e., eq 1) in which the reversible step 5 is followed by the following sequence of steps.8 p =

k6

C0(CN)c4- +Co(CN),'-

+ CN-

(6)

fast

Co(CN)r3-

+ CO +C O ( C N ) ~ C O ~ -

(8)

+

7.0 6.9 6.0 7.6 6.6~ 6.7 6.0 6.2 6.1 6.0d 5.9 6.4 6.6 2.2 13.8

6.2 6.6 2.1 13.4

fast + CO + C O ( C N ) ~ ( C O ) ~+~ -CN-

(7)

(8) It is likely that reaction 8 actually proceeds by a dissociative stepwise mechanism analogous to that depicted by eq 6 and 7, but our results provide no direct information about this.

According to this interpretation, k = k5k6/(k-s k 6 ) . Combining the present value of k with the previously ~ determined values of ks (1.6 X lo-' M-I ~ e c - l )and k-5 (1.1 X los ~ e c - ' )yields ~ k-s/k6 = 25 and ks = 4.5 X IO3 sec-'. The proposed mechanistic scheme, in addition to explaining the observed kinetic patterns, seems highly plausible on chemical grounds. The dissociation of C O ( C N ) ~ seems ~reasonable in view of the known tendency of five-coordinate ds complexes such as Ni(CN)53(with which CO(CN),~- is isoelectronic) to undergo such dissociation in aqueous solution. Since Co(CN)s4- possesses an 18-electron valence shell, it also seems highly likely that the substitution of CN- by CO to form C O ( C N ) ~ C O ~will - proceed through a dissociative mechanism (Le., eq 6 and 7) rather than an associative one. Failure to observe any inhibition of the reaction by CN- up to 0.25 M CN- (compared to the much lower M> implies prevailing CO concentration of only ~ -form CO(CN)~that the addition of CO to C O ( C N ) ~ to C03- (eq 7) is much faster than the corresponding addition of CN- to form C O ( C N ) ~ ~ -i.e., , to the reverse of eq 6. This is not surprising since CO, being a better T acceptor than CN-, is expected to have a greater tendency to favor five-coordination relative to four-coordination for low-spin ds complexes. This difference is likely to be particularly pronounced in the case of a highly negatively charged complex such as C O ( C N ) ~ ~ - . The occurrence of reaction 6, in competition with the reverse of reaction 5 , has also been inferred recently by Lim and Ansonlo from observations on the electrochemical reductions of cobalt cyanide complexes. Acknowledgment. Support of this work through grants from the National Institutes of Health and the National Science Foundation is gratefully acknowledged. One of us (M. P.) also thanks the U. S . Government Fulbright-Hays Program for a travel grant and the University of Zagreb for a leave of absence. (9) G. D. Venerable I1 and J. Halpern, ibid., 93, 2176 (1971). (10) H. S. Lim and F. C . Anson, J . Electround. Chem., 31, 297 (1971).

Guastalla, Halpern, Pribani; 1 Carbonylation of HydridopentacyanocobaItate(III)